Cancer immunosurveillance is a potential mechanism for immune recognition and elimination of pre-emergent tumor cells, but little is known about the tumor cell-derived danger signals that drive immunosurveillance. Genetically engineered mouse models (GEMMs) recapitulate cardinal features of human cancers, including interactions between tumor cells and their microenvironment, and have helped uncover complexities underlying tumor development. We have previously used GEMMs to identify similarities and differences between lung adenocarcinoma and sarcoma development. Expression of oncogenic KrasG12D by muscle cells activates a genetic circuit involving p19 (Arf) and p53 that promotes arrest and stops tumor development. By contrast, lung tumors bypass this genetic circuit by epigenetically silencing p19 (Arf). In advanced tumor cells, signals down-stream of p53 also trigger arrest and mediate elimination by NK and other innate immune cells. The fate of KrasG12D-arrested muscle cells is unclear, but immunosurveillance and elimination by innate immune cells could reduce the chances these pre-neoplastic cells acquire additional genetic alterations and develop into sarcomas. In Aim 1 of this application, we will directly investigate the role of innate immune cells in detecting and clearing pre-neoplastic cells in a mouse model in which the events controlling the initial acquisition of oncogenic K-rasG12D are independent of those controlling p53 deletion (required for release from an arrested state). Additionally, we will test whether NK cells can augment and/or inhibit clearance by expressing activating and inhibitory ligands in these lesions. The comparison of KrasG12D expressing p53-deficient sarcomas and lung adenocarcinomas provides a powerful platform for investigating how early events in tumorigenesis determine whether anti-tumor T cell responses are protective in vivo. Despite tumor development in identical mice, with identical initiating genetic alterations, expressing identical tumor antigens, and eliciting responses from identical populations of naive T cells, antigen-expressing lung adenocarcinomas develop in mice while antigen- expressing sarcomas do not. In Aim 2, we will compare early innate immune responses to sarcomas and lung adenocarcinomas and determine how environmental and cell-intrinsic responses to transformation contribute to innate activation and anti-tumor T cell responses. In Aim 3, we will identify the mechanisms by which immune cells become alerted to the presence of pre-neoplastic and fully transformed tumor cells and whether knockdown of danger signals can help pre-emergent tumors subvert immunosurveillance mechanisms and grow in vivo. These Aims will strengthen our understanding of the molecular and cellular events that regulate immunosurveillance at the earliest stages of malignant transformation and help us to understand why certain tumor types are more immunogenic as a result.